zai-org / glm4_moe_lite

GLM-4.7-Flash

Name: GLM-4.7-Flash architecture and memory requirements
Creator: LLM Bottleneck
License: https://creativecommons.org/licenses/by/4.0/

31.2 billion parameters.

Open in the calculator →

Architecturepublished data

Quick answer

GLM-4.7-Flash needs about 20.4 GB at Q4_K_M — the format most people download — with 8,192 tokens of context. It fits entirely on 12 of 18 common devices, starting with the NVIDIA GeForce RTX 4090 (24 GB). What else fits in 24 GB. No hardware for it? Rent a GPU that runs it, priced live.

Smallest device that fitsNVIDIA GeForce RTX 409024 GB · 1008 GB/s~35 tok/s Faster than you readOpen in the calculator →
Fastest in this listNVIDIA GeForce RTX 509032 GB · 1792 GB/s~62 tok/s Faster than you readOpen in the calculator →

Sized at 8,192 tokens of context with the whole model in device memory, across 18 common devices. Speeds are estimates, not benchmarks.

Find the NVIDIA GeForce RTX 4090, the smallest card listed here that holds it: Amazon ↗ · eBay (new and used) ↗Store links may pay us a commission. They never decide which card is suggested — the memory arithmetic does.

Weight-file size by format

every format, with its evidence

Format	Size	Range	Basis
FP16	62.5 GB	43.7 GB – 63.1 GB	size range
Q8_0	33.2 GB	23.2 GB – 34.1 GB	size range
Q6_K	25.6 GB	17.9 GB – 26.6 GB	size range
Q5_K_M	22.3 GB	15.0 GB – 26.6 GB	size range
Q5_0	21.8 GB	15.0 GB – 26.6 GB	size range
Q4_K_M	19.2 GB	12.3 GB – 26.6 GB	size range
Q4_0	18.2 GB	12.3 GB – 26.6 GB	size range
Q3_K_M	15.6 GB	9.39 GB – 26.6 GB	size range
Q2_K	12.4 GB	7.17 GB – 26.6 GB	size range

A file size is not the memory a run needs: the KV cache and the runtime reserve come on top, and the calculator adds both.

What the labels mean

published data: Read from a published source — a file's byte count, a model's configuration or a manufacturer's specification — or exact arithmetic on such values. Not a measurement on a machine.
reconstructed size: Weight size reconstructed from the pinned architecture, because no published file exists.
size range: Only a lower and an upper bound are claimed for this weight size.

Hardware ladder

Q4_K_M at 8,192 tokens · 12 of 18 devices hold it entirely

Device	Memory	Verdict	Decode	Calculator
NVIDIA GeForce RTX 5070	20.4 GB of 12 GB	21 layers on system RAM	~6.0 tok/s with offload	Open →
AMD Radeon™ RX 9070 XT	20.4 GB of 16 GB	11 layers on system RAM	~9.5 tok/s with offload	Open →
NVIDIA GeForce RTX 4080	20.4 GB of 16 GB	11 layers on system RAM	~9.4 tok/s with offload	Open →
NVIDIA GeForce RTX 5080	20.4 GB of 16 GB	11 layers on system RAM	~10 tok/s with offload	Open →
NVIDIA GeForce RTX 5070 Ti	20.4 GB of 16 GB	11 layers on system RAM	~10 tok/s with offload	Open →
NVIDIA GeForce RTX 5060 Ti 16GB	20.4 GB of 16 GB	11 layers on system RAM	~8.0 tok/s with offload	Open →
NVIDIA GeForce RTX 4090	20.4 GB of 24 GB	fits	~35 tok/s	Open →
AMD Radeon™ RX 7900 XTX	20.4 GB of 24 GB	fits	~38 tok/s	Open →
NVIDIA GeForce RTX 3090	20.4 GB of 24 GB	fits	~32 tok/s	Open →
NVIDIA GeForce RTX 5090	20.4 GB of 32 GB	fits	~62 tok/s	Open →
NVIDIA RTX 6000 Ada Generation	20.4 GB of 48 GB	fits	~33 tok/s	Open →
Apple M4 Pro	20.4 GB of 64 GB	fits	~12 tok/sAbout reading pace	Open →
Apple M5 Pro	20.4 GB of 64 GB	fits	~14 tok/sAbout reading pace	Open →
AMD Ryzen AI Max+ 395 with Radeon 8060S	20.4 GB of 128 GB	fits	~10 tok/sAbout reading pace	Open →
Apple M3 Max	20.4 GB of 128 GB	fits	~17 tok/sAbout reading pace	Open →
Apple M4 Max	20.4 GB of 128 GB	fits	~23 tok/sAbout reading pace	Open →
NVIDIA DGX Spark	20.4 GB of 128 GB	fits	~9.5 tok/sSlow	Open →
Apple M2 Ultra	20.4 GB of 192 GB	fits	~31 tok/s	Open →

Decode is a calibrated estimate from the published calibration; capacity is the manufacturer's published ceiling, not guaranteed free memory. Fitting in memory is not the same as loading: whether the runtime and version you have supports this architecture and format on that machine has not been tested here. Offloaded rows assume there is enough system RAM for the overflow — the calculator checks that against the RAM you declare. Each “Open” link keeps this model, format and context.

Run it in the cloud

Rent a GPU that runs GLM-4.7-Flash

Every rentable machine that holds the whole model, from Vast.ai and RunPod, with this site's speed estimate for each and the provider's own price, read live. Cheapest first; sort by cost per token or speed instead, or switch the format.

Prices · 18:30 UTC, 22 Sept

Machine	Speed	Per hour	Per M tokens	Where	Rent
2× RTX 306024 GBcheapest	~24 tok/sAbout reading pace	$0.11	$1.24	Vast.aiMarketplace · 98.0% reliable	Rent on Vast.ai
1× RTX 309024 GBrecommendedbest value	~32 tok/sFaster than you read	$0.12	$1.05	Vast.aiMarketplace · 99.5% reliableRunPod $0.50/h	Rent on Vast.ai
1× RTX 3090 Ti24 GB	~34 tok/sFaster than you read	$0.19	$1.49	Vast.aiMarketplace · 99.6% reliableRunPod $0.27/h	Rent on Vast.ai
2× RTX 407024 GB	~34 tok/sFaster than you read	$0.19	$1.54	Vast.aiMarketplace · 99.8% reliable	Rent on Vast.ai
2× RTX 4060 Ti 16GB32 GB	~19 tok/sAbout reading pace	$0.22	$3.20	Vast.aiMarketplace · 99.7% reliable	Rent on Vast.ai
1× L424 GB	~10 tok/sAbout reading pace	$0.26	$6.83	Vast.aiMarketplace · 98.9% reliableRunPod $0.49/h	Rent on Vast.ai
2× RTX 5060 Ti 16GB32 GB	~30 tok/sFaster than you read	$0.26	$2.34	Vast.aiMarketplace · 99.5% reliable	Rent on Vast.ai
1× RTX 6000 Ada48 GB	~33 tok/sFaster than you read	$0.26	$2.18	Vast.aiMarketplace · 99.9% reliableRunPod $0.84/h	Rent on Vast.ai

No GPU at all

Use GLM-4.7-Flash by the token

Novita AI$0.40 / M output · $0.070 / M inputOpen on Novita AI
OpenRouter$0.40 / M output · $0.061 / M inputOpen on OpenRouter

For one person chatting, that is about 2.6× cheaper than the best-value rented card above ($1.05 per million tokens). A rented GPU pays off when you keep it busy — many requests batched together, long agent runs — or when the data must stay on a machine you control, or the exact file you want is not served anywhere.

First time renting a GPU? How it works, in four steps

Create an account and add credit. Both providers are prepaid: $10 is enough to try any card on this page for hours. A new RunPod account that arrives through a referral link gets a one-time $5 credit when it first adds $10.
Pick the card and the count shown here, and a template: llama.cpp or vLLM for an OpenAI-compatible API, or one with a web chat if you only want to talk to the model. “How to launch it” below gives the exact image and command.
Wait for the download. The weights are fetched on the machine; tens of gigabytes take a few minutes on a data-centre connection. You pay by the second from the moment it starts.
Stop it when you are done. A running machine bills even when idle. A stopped one costs nothing per hour, though its disk is usually still billed until you delete it.

How to launch it

No Q4_K_M file of GLM-4.7-Flash is catalogued here, so there is no exact command to vouch for. Start the machine from the llama.cpp server image ghcr.io/ggml-org/llama.cpp:server-cuda and point -hf at a Q4_K_M build from the publisher or a quantizer you trust — search Hugging Face for one. Check its file size against the memory figure on this page before you rent.

Every machine holds the whole model at Q4_K_M and 8,192 tokens of context, no offload. Speeds are this site’s single-stream decode estimates; prices are what each provider’s own API quoted, on-demand, for the whole machine. Vast hosts below 98% measured reliability are left out.

Referral links Vast.ai, RunPod and Novita pay us a share of what you spend if you sign up through these buttons. It costs you nothing, and it never decides an order or a recommendation: both are computed from the live price and the speed, and options that pay us nothing are listed and recommended on the same terms. How we rank

Buying a card instead, or paying by the token? Compare a year of GLM-4.7-Flash three ways →

Under the hood

Architecture, read from the publisher’s file

The numbers every figure above is computed from, with the file they came from.

Architecture

architecture.json · pinned 7dd20894a642

✓ Architecture read from the published config.json

Retrieved 2026-09-01 at pinned commit 7dd20894a642; parameter count from the safetensors index at the same revision.

Architecture: glm4_moe_lite
Layers: 47
Hidden size: 2,048
Attention heads: 20
KV heads: 20
Feed-forward width: 10,240
Vocabulary: 154,880
Context ceiling: 202,752
RoPE theta: 1,000,000
Experts: 64
Experts / token: 4
Expert width: 1,536
Shared experts: 1
Latent KV rank: 512

Open the exact official config.json ↗

GLM-4.7-Flash, card by card

One page per device: whether GLM-4.7-Flash fits, at which formats, and how fast.

All 27Fewer

More from zai-org

GLM-4.5 GLM-4.5-Air GLM-4.5V GLM-4.6 GLM-4.6V GLM-4.6V-Flash GLM-4.7 GLM-5

Use this answer in your own product

As JSON, from the API

curl -s https://llmbottleneck.com/v1/analyze \
  -H "Authorization: Bearer $LLMB_KEY" -H "content-type: application/json" \
  -d '{"model":"zai-org-glm-4-7-flash","quantization":"Q4_K_M","context":8192,"hardware":"nvidia-geforce-rtx-4090"}'

Same engine, same evidence, every field sourced. Free key in one step →

On your page, as a widget

<script src="https://llmbottleneck.com/widget.js"
  data-model="zai-org-glm-4-7-flash" data-quantization="Q4_K_M"
  data-hardware="nvidia-geforce-rtx-4090" data-context="8192"></script>

No key needed. Unbranded, with your own buy button, on Pro and Business →

Citing this page

LLM Bottleneck. “GLM-4.7-Flash VRAM and hardware requirements.” Architecture from zai-org/GLM-4.7-Flash at revision 7dd20894a642, retrieved 2026-09-01. https://llmbottleneck.com/models/zai-org-glm-4-7-flash

Every figure above is either the published value or a reconstruction whose measured error is on the accuracy page.